A basic thesis of Neokantian epistemology and philosophy of science contends that the knowing subject and the object to be known are only abstractions. What really exists, is the relation between both. For the elucidation of this "knowledge relation" ("Erkenntnisrelation") the Neokantians of the Marburg school used a variety of mathematical metaphors. In this contribution I"d like to reconsider some of these metaphors proposed by Paul Natorp one of the leading members of the Marburg school. It is shown that Natorp's metaphors are not unrelated to those used in some currents of contemporary epistemology and philosophy of science.